Table 7.

Results of within-domain and pair-wise cross-domain support vector regression on valence observer ratings for sound (emotional sound database), music (NTWICM database), and spontaneous and enacted speech (VAM/GEMEP databases).

r	Test on				Mean
Train on	Sound	Music	Speech
			Sp.	En.
(A) FULL FEATURE SET
Sound	0.40**	−0.11**	0.21**	−0.02−	0.12
Music	−0.17 °	0.80**	−0.13*	0.08−	0.15
Speech/Sp.	0.11−	−0.15**	0.46**	0.21−	0.16
Speech/En.	−0.06−	−0.18**	0.12*	0.26o	0.03
Mean	0.07	0.09	0.17	0.13	0.12
(B) 200 TASK-SPECIFIC FEATURES
Sound	0.51**	0.36**	0.27**	0.48**	0.41
Music	0.40**	0.82**	0.33**	0.52**	0.52
Speech/Sp.	0.30**	0.45**	0.44**	0.26o	0.36
Speech/En.	0.45**	0.60**	0.36**	0.50**	0.48
Mean	0.41	0.56	0.35	0.44	0.44
(C) 200 GENERIC FEATURES
Sound	0.26**	0.41**	0.27**	0.12−	0.27
Music	0.27**	0.75**	0.33**	0.25o	0.40
Speech/Sp.	0.20*	0.45**	0.35**	0.19−	0.30
Speech/En.	0.20**	0.44**	0.32**	0.23−	0.30
Mean	0.23	0.52	0.32	0.20	0.32

Significance denoted by **p < 0.001, *p < 0.01, °p < 0.05, ⁻p ≥ 0.05; Bonferroni corrected p-values from two-sided paired sample t-tests. Full ComParE feature set (cf. Tables 2 and 3); 200 top features selected by CDCC² for specific within-domain or cross-domain regression tasks; Generic features: 200 features selected by CDCC³ across sound, music, and speech domains (cf. Table 5).